Anthropic removed a hidden tracker from Claude Code that secretly monitored Chinese users after security researcher Thereallo exposed the code last week. The tracker used prompt steganography to collect user timezone, proxy, and potential connection data to Chinese AI labs, sending information to Anthropic that most users wouldn't detect. Thereallo condemned the tracking as a serious breach of user trust. Anthropic engineer Thariq Shihipar confirmed on X that the code was added as an experiment in March to prevent account abuse from unauthorized resellers and protect against distillation attacks. The removal came after public exposure, with Shihipar stating engineers had developed stronger mitigations since the tracker's implementation.
Security Researcher Exposes Hidden Tracking Code in Claude Code
Last week, web developer Thereallo discovered the hidden code while researching privacy issues in Claude Code. The researcher found Anthropic was using prompt steganography to hide tracking code in plain sight. The code relied on shorthand markers to quietly flag users' timezone, proxy settings, and potential connections to Chinese AI labs that Anthropic has accused of distillation attacks. While not malicious, the code sent information to Anthropic without most users' knowledge or detection.
Anthropic Engineer Confirms March Experiment Targeting Unauthorized Resellers
Thariq Shihipar stated on X that the tracker was added to Claude Code as an experiment in March. According to Shihipar, the code was meant to prevent account abuse from unauthorized resellers and protect against distillation. The Washington Post found unauthorized retailers have sold access to free Claude models for $1 a month, while pro subscriptions that can cost $100 monthly sell for as little as $12. Shihipar said Anthropic had been meaning to take down the hidden code for a while because engineers landed stronger mitigations since then.
Privacy Advocates Criticize Surveillance Amid US Government Access Refusal
Privacy advocates criticized the explanation, warning the code is evidence Anthropic is willing to cross lines to surveil users. The criticism comes as Anthropic previously refused to allow the US government to use Claude to surveil US users, riling the Trump administration. The AI firm has since sued the White House over the clash.
FAQ
What did Anthropic remove from Claude Code last week?
Anthropic removed a hidden tracker that secretly monitored Chinese users in Claude Code after security researcher Thereallo exposed the code. The tracker used prompt steganography to collect user timezone, proxy, and potential connection data to Chinese AI labs.
Why did Anthropic add the tracking code to Claude Code?
Anthropic engineer Thariq Shihipar confirmed the code was added as an experiment in March to prevent account abuse from unauthorized resellers and protect against distillation attacks. Unauthorized retailers have sold Claude access at significantly reduced prices compared to official subscriptions.
How did privacy advocates react to the hidden tracker?
Privacy advocates criticized Anthropic, warning the code is evidence the company is willing to cross lines to surveil users. The criticism is notable given Anthropic previously refused to allow the US government to use Claude to surveil US users.
